Abstract
A classification infrastructure built upon Discriminant Analysis (DA) has been developed at NorthWest Research Associates for examining the statistical differences between samples of two known populations. Originating to examine the physical differences between flare-quiet and flare-imminent solar active regions, we describe herein some details of the infrastructure including: parametrization of large datasets, schemes for handling “null” and “bad” data in multi-parameter analysis, application of non-parametric multi-dimensional DA, an extension through Bayes’ theorem to probabilistic classification, and methods invoked for evaluating classifier success. The classifier infrastructure is applicable to a wide range of scientific questions in solar physics. We demonstrate its application to the question of distinguishing flare-imminent from flare-quiet solar active regions, updating results from the original publications that were based on different data and much smaller sample sizes. Finally, as a demonstration of “Research to Operations” efforts in the space-weather forecasting context, we present the Discriminant Analysis Flare Forecasting System (DAFFS), a near-real-time operationally-running solar flare forecasting tool that was developed from the research-directed infrastructure.
Highlights
The prospect of forecasting rare events such as solar flares is a daunting one, especially in situations where the exact trigger mechanism or threshold for instability is not yet known
The full derivations and descriptions can be found in the cited references; of note here are the critical functions of each: the Hanssen and Kuipers discriminant (H&KSS) measures the discrimination between the Probability of Detection and the False Alarm Rate, the Appleman skill score (ApSS) measures the skill against the climatological forecast, and the Brier Skill Score (BSS) evaluates the performance of probabilistic forecasts against observed occurrence
An investigative infrastructure which has been developed at NorthWest Research Associates based on Discriminant Analysis classifiers has been described, and briefly demonstrated in the context of research centered on distinguishing flare-ready from flare-quiet solar active regions
Summary
The prospect of forecasting rare events such as solar flares is a daunting one, especially in situations where the exact trigger mechanism or threshold for instability is not yet known. Different classes of active region (based on size and sunspot-group characteristics) were observed to produce flares at different rates, and applying Poisson statistics resulted in probabilistic forecasts for flares (McIntosh, 1990; Bornmann & Shaw, 1994). This approach forms the basis for many forecasts published today, including those from the US National Oceanic and Atmospheric Administration (NOAA)/ Space Weather Prediction Center (SWPC) (Sawyer et al, 1986; Gallagher et al, 2002; Murray et al, 2017; Steenburgh & Balch, 2017). It is presently in use to aid target selection for the Hinode mission (Kosugi et al, 2007), its limited field-of-view instruments (the Solar Optical Telescope, Tsuneta et al, 2008, and the EUV Imaging Spectrograph, Culhane et al, 2007)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have